1. Field of the Invention
Methods and apparatuses consistent with the present invention relate to a digital broadcasting program, and more particularly, to a method and apparatus for segmenting a recorded news program according to topics.
2. Description of the Related Art
Recently, electronic devices such as a digital televisions (TVs) or personal video recorders (PVRs) having large storage capacities have been introduced so that users can easily record a broadcasting program via a digital broadcasting channel and can conveniently reproduce and watch the broadcasting program at a desired time.
However, in the case of a news program, a plurality of articles are generally included in one news program. Thus, even though a user has recorded a news program, the user has to inconveniently search for a desired topic in the recorded news program by executing a fast forward function.
In order to solve such inconveniences, research has been conducted, aiming to divide a recorded news program into segments by analyzing a caption, voice, an image, and the like. However, the research depending on the analysis of the caption, the voice, and the image has problems in that a border of a research field is ambiguous, related technologies are complicated, and accuracy deteriorates.
For example, accuracy of a segmentation method according to voice analysis, or according to image analysis particularly deteriorates when more than two announcers or reporters report a same topic. Also, a segmentation method according to caption analysis analyzes words included in a caption, and detects an issue change, thereby forming segments. However, not every news program provides the caption, and moreover, since the caption is typed in real-time by a shorthand writer who listens to voice from the news program, thus, there may be many typographic errors. Hence, accuracy of the segmentation method corresponding to such a caption is also very low.